Cover versus biomass as an estimate of species abundance:: does it make a difference to the conclusions?

Different measures of species abundance, particularly cover and biomass, are often used as if they were interchangeable. We ask to what extent the measure of abundance affects conclusions reached in plant community ecology. The test data comprised 35 plots, each 1 m x 1 m, on serpentine vegetation in Tuscany, Italy, in which both cover and biomass were measured. The measure of abundance used made rather little difference when plots were examined in terms of the rank abundance of the species present, and little difference to the results of a DCA-ordination. Likewise, species showed similar variation across plots in cover as in biomass. In calculating species diversity, agreement between measures of abundance was lower, but still good. For evenness, the correlation between cover and biomass results was lower again, though more so with some indices than with others. For the shape of the RAD (Relative Abundance Distribution), cover and biomass gave quite different results, the Zipf-Mandelbrot RAD model most commonly giving the best fit to the cover data, but the General lognormal model to the biomass data. These tendencies can be related to the intrinsic characteristics of cover and biomass as measures of abundance. The parameters of the model fits also differed between abundance measures. We conclude that the importance of using biomass as the measure of abundance depends on the ecological feature being examined. The difference between the two measures may be greater in other vegetation; very little information is available. In the current state of knowledge it is dangerous to use another measure as a surrogate for biomass.